RESEARCH ARTICLE Open Access

Resolved phylogeny and biogeography ofthe root pathogen Armillaria and itsgasteroid relative, GuyanagasterRachel A. Koch1, Andrew W. Wilson2, Olivier Séné3, Terry W. Henkel4 and M. Catherine Aime1*

Abstract

Background: Armillaria is a globally distributed mushroom-forming genus composed primarily of plant pathogens.Species in this genus are prolific producers of rhizomorphs, or vegetative structures, which, when found, are oftenassociated with infection. Because of their importance as plant pathogens, understanding the evolutionary originsof this genus and how it gained a worldwide distribution is of interest. The first gasteroid fungus with close affinitiesto Armillaria—Guyanagaster necrorhizus—was described from the Neotropical rainforests of Guyana. In this study, weconducted phylogenetic analyses to fully resolve the relationship of G. necrorhizus with Armillaria. Data sets containingGuyanagaster from two collecting localities, along with a global sampling of 21 Armillaria species—including newlycollected specimens from Guyana and Africa—at six loci (28S, EF1α, RPB2, TUB, actin-1 and gpd) were used. Threeloci—28S, EF1α and RPB2—were analyzed in a partitioned nucleotide data set to infer divergence dates and ancestralrange estimations for well-supported, monophyletic lineages.

Results: The six-locus phylogenetic analysis resolves Guyanagaster as the earliest diverging lineage in the armillarioidclade. The next lineage to diverge is that composed of species in Armillaria subgenus Desarmillaria. This subgenus iselevated to genus level to accommodate the exannulate mushroom-forming armillarioid species. The final lineage todiverge is that composed of annulate mushroom-forming armillarioid species, in what is now Armillaria sensu stricto. Themolecular clock analysis and ancestral range estimation suggest the most recent common ancestor to the armillarioidlineage arose 51 million years ago in Eurasia. A new species, Guyanagaster lucianii sp. nov. from Guyana, is described.

Conclusions: The armillarioid lineage evolved in Eurasia during the height of tropical rainforest expansion about 51million years ago, a time marked by a warm and wet global climate. Species of Guyanagaster and Desarmillaria representextant taxa of these early diverging lineages. Desarmillaria represents an armillarioid lineage that was likely much morewidespread in the past. Guyanagaster likely evolved from a gilled mushroom ancestor and could represent a highlyspecialized endemic in the Guiana Shield. Armillaria species represent those that evolved after the shift in climate fromwarm and tropical to cool and arid during the late Eocene. No species in either Desarmillaria or Guyanagaster are knownto produce melanized rhizomorphs in nature, whereas almost all Armillaria species are known to produce them. Theproduction of rhizomorphs is an adaptation to harsh environments, and could be a driver of diversification in Armillaria byconferring a competitive advantage to the species that produce them.

Keywords: Armillaria root rot, Cameroon, Eocene, Fungal taxonomy, Gasteromycetation, Guiana Shield, Melanin,Mushroom evolution, Physalacriaceae, Systematics

* Correspondence: maime@purdue.edu1Department of Botany and Plant Pathology, Purdue University, WestLafayette, IN 47907, USAFull list of author information is available at the end of the article

© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Koch et al. BMC Evolutionary Biology (2017) 17:33 DOI 10.1186/s12862-017-0877-3

http://crossmark.crossref.org/dialog/?doi=10.1186/s12862-017-0877-3&domain=pdfmailto:maime@purdue.eduhttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/publicdomain/zero/1.0/

BackgroundThe globally distributed mushroom-forming genus Armil-laria contains species that are frequently encountered astree pathogens in natural forests [1, 2] as well as silvicul-tural and agronomic systems [3, 4], and are the causal agentof the root disease Armillaria root rot [5]. Besides beingpathogens, Armillaria species also play a critical role asdecomposers. Extensive infection observed in stumps androots and long possession of the substrate suggest thatArmillaria species contribute significantly to decompos-ition and mineral cycling within many forests [6].Armillaria species form basidiomes—in this case,

mushrooms—which serve as reproductive structuresthat fruit seasonally when conditions are optimal. Eachof these basidiomes produces their reproductive propa-gules, or basidiospores, on an exposed spore-bearingsurface, or hymenium. At maturity, the basidiospores arepropelled into the air column via a mechanism offorcible spore discharge known as ballistospory. Basidio-spores of other fungi capable of ballistospory have beenshown to disperse between continents [7–9]. Based onmolecular clock analyses, the origin of Armillaria post-dates the Gondwana break-up, leading to the hypothesisthat several such long distance basidiospore dispersalevents led to the global distribution of Armillaria [10].One change in basidiome form that has occurred re-

peatedly within mushroom-forming lineages is the shiftfrom producing basidiospores on an exposed hymeniumto producing them within an enclosed hymenium. Thisloss of an exposed hymenium has resulted in the loss ofballistospory [11]. Fungi that undergo this change willhereinafter be referred to as gasteroid fungi as they gothrough a process called gasteromycetation. A suite ofchanges in basidiome morphology, including the devel-opment of a fully enclosed basidiospore-bearing mass(termed a gleba) that is encased in a specialized covering(termed a peridium), typically take place during gastero-mycetation. Gasteromycetation is believed to be a uni-directional process, as gasteroid fungi have never beenobserved to regain the mechanism of ballistospory [11].In order to persist, gasteroid fungi have evolved a diversearray of dispersal mechanisms that engage exogenousforces. Examples include animal mycophagy for volatile-producing underground truffle-like basidiomycetes [12]and the “bellows” mechanism of puffballs [13].Gasteroid fungi have evolved independently many

times from ballistosporic ancestors within the Agarico-mycetidae (e.g. [14–20]). These ancestors of gasteroidfungi include agaricoid lineages (i.e. those with lamellatehymenophores in the Agaricales and Russulales) andboletoid lineages (i.e. those with porose-tubulose hyme-nophores in the Boletales). To date, gasteroid fungi haveonly been documented as ectomycorrhizal (ECM) orsaprotrophic in their ecology (e.g. [11, 21–26]).

The first known gasteroid fungus closely related toArmillaria was described in 2010 [27]. The monotypicgenus, Guyanagaster T. W. Henkel, Aime & M. E. Sm.and the species G. necrorhizus T. W. Henkel, Aime & M.E. Sm. (Basidiomycota; Agaricales; Physalacriaceae) areonly known from the Pakaraima Mountains of Guyanain the Guiana Shield region of South America [27]. Thisfungus produces subhypogeous basidiomes that are oftenattached to decaying tree roots, has a thick black peridiumwith pyramidal warts, a reduced stipe, and a tough, gelat-inous gleba that changes from white to pink to brick redwith maturation [27]. This species is remarkable in thesense that it does not possess the morphological hallmarksof gasteroid fungi adapted for any of the aforementioneddispersal mechanisms (e.g. mammal or mechanicaldispersal).To our knowledge, until the discovery of Guyanagaster,

no known gasteroid fungus has been demonstrated tohave evolved from a pathogenic lineage. Guyanagasternecrorhizus basidiomes are almost always found fruitingfrom dead and decaying tree roots, in this case primarilyof the forest-dominating trees in the genus DicymbeSpruce ex Benth. (Fabaceae subfam. Caesalpinioideae).Roots to which G. necrorhizus are attached display signsof white rot, indicating a wood decay capability for thisfungus. Given the association of G. necrorhizus to deadand decaying roots and its close relationship to Armillariaspecies, it is conjectured that it is also has pathogenic cap-abilities, but no experimental evidence exists to confirmthis.Prior multi-gene phylogenetic analyses by Henkel et al.

were equivocal in that some loci suggested G. necrorhizuswas sister to Armillaria, while other loci showed G.necrorhizus derived from within Armillaria [27]. Resolvingthe phylogenetic placement of G. necrorhizus in relationto Armillaria will create a more complete picture ofarmillariod (a term to denote both Armillaria andGuyanagaster species) evolution through time andallow us to understand traits that led to the success ofthis group. A resolved phylogeny will also give us a ro-bust framework in which to understand the evolutionof Guyanagaster.Here we use molecular data from G. necrorhizus and

Armillaria species from every major infrageneric lineage,including Armillaria puiggarii Speg.—the only Armil-laria species to be collected in the same region as G.necrorhizus—to provide a fully resolved phylogenetichypothesis for armillarioid fungi. We also erect a newgenus, Desarmillaria, composed of the exannulatemushroom-forming armillarioid species, to accommodatethose species formerly placed in Armillaria subgenusDesarmillaria. A second species of Guyanagaster, Guya-nagaster lucianii, is described as new to science. Usingthese data we produce a time-calibrated phylogeny, which

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 2 of 16

we analyze in combination with ancestral range esti-mations, in order to examine the morphological andbiogeographic evolution of armillarioid fungi.

MethodsCollecting and morphological analysesCollecting expeditions to the Upper Potaro River Basin(hereafter referred to as the Potaro) in the west-centralPakaraima Mountains of Guyana were conducted dur-ing the rainy seasons of May–July of 2002–2013. Fungiwere collected within a 15-km radius of a previouslyestablished base camp (5°18'04.80"N, 59°54'40.40"W) inforests dominated by Dicymbe corymbosa Spruce exBenth. and Dicymbe altsonii Sandw. (Fabaceae subfam.Caesalpinioideae) [28, 29]. In December 2013–January2014, a collecting expedition to the Mabura EcologicalReserve (hereafter referred to as Mabura) in MaburaHill, Guyana—approximately 125 km from the Potar-o—was conducted. Fungi were collected within a 3-kmradius of the field station (5°10'09.26"N, 58°42'14.40"W)in forests with high densities of D. altsonii [30]. Newcollections of African Armillaria species were made inthe Dja Biosphere Reserve in the East Province ofCameroon during the rainy season of August–September2014 within a 2-km radius of the Dja base camp (3°21'29.80"N, 12°43'46.90"W).Fresh morphological characteristics and substratum

relationships were described in the field. Color was de-scribed subjectively and coded according to Kornerup andWanscher [31], with color plates noted in parentheses.The descriptions of the collected Armillaria specimenswere compared to the descriptions of validly publishedspecies. Additionally, attempts to obtain specimen cul-tures were made in the field by placing small pieces ofunexposed basidiome tissue on Potato Dextrose Agar(PDA) (BD Difco™, Franklin Lakes, New Jersey, USA) andgrown for one month at room temperature. Specimenswere then field-dried with silica gel.Micromorphological features on dried specimens were

examined in the laboratory using an Olympus BH2-RFCA compound microscope. Two dried Armillariaspecimens from Africa, TH 9926 and THDJA 91, andfour dried Guyanagaster specimens, MCA 4424, RAK84, RAK 88 and RAK 89, were rehydrated in 70% etha-nol and then sectioned by hand and mounted in water,Melzer’s reagent and 1% aqueous Congo red. Twentyrandomly selected basidiospores were measured per col-lection under a 100× objective. Length/width Q valuesfor basidiospores are reported as Qr (range of Q valuesover 20 basidiospores measured) and Qm (mean of Qvalues ± SD). Specimens were deposited in the followingherbaria: PUL (Kriebel Herbarium), BRG (GuyanaNational Herbarium), HSC (Humboldt State University)and YA (National Herbarium of Cameroon).

Molecular methodsDNA was extracted from basidiome tissue using theWizard® Genomic DNA Purification kit (Promega Co.,Madison, Wisconsin, USA). Specimens housed in theKriebel Herbarium at Purdue University were used toobtain DNA if sequence data for equivalent taxa werenot available on the NCBI Nucleotide database. PCR re-actions included 12.5 μL of MeanGreen 2× Taq DNAPolymerase PCR Master Mix (Syzygy Biotech, GrandRapids, Michigan, USA), 1.25 μL of each primer (at10 μM) and approximately 100 ng of DNA. The finalPCR reaction volume was 25 μL. The recommendedcycling conditions for each primer pair we used werefollowed.To determine the identity of the recently collected

Armillaria specimens from Guyana and Africa, as wellas to confirm the identity of the recently collectedGuyanagaster specimens, PCR was performed to ac-quire sequence data from the internal transcribed spa-cer (ITS) region (inclusive of ITS1, 5.8S and ITS2regions), using the primer pair ITS1F/ITS4B [32]. Toanalyze the phylogenetic relationship of the armillarioidfungi, PCR was performed on all specimens at the threefollowing loci: nuclear ribosomal large subunit DNA(28S), Elongation Factor 1-α (EF1α) and RNA polymer-ase II (RPB2) genes using the following primer pairs,respectively: LROR/LR6 [33, 34], 987 F/2218R [35] andbRPB2-6 F/bRPB2-7.1R [36].Uncleaned PCR products were sent to Beckman

Coulter, Inc. (Danvers, Massachusetts, USA) for sequen-cing. Sequences were manually edited using Sequencher5.2.3 (Gene Codes Corporation, Ann Arbor, Michigan,USA). The ITS sequences we generated were used asqueries in the sequence similarity search tool, NCBIBLAST, to search the NCBI Nucleotide database.

Phylogenetic analysesIn order to resolve the phylogeny of Guyanagaster andArmillaria, we compiled a dataset composed ofsequence data from both Armillaria and Guyanagasterspecimens. Two closely related species in the Physala-criaceae, Oudemansiella mucida and Strobilurus esculen-tus, served as outgroup taxa. Six loci were used (28S,EF1α, RPB2, Actin-1 (actin-1), Glyceraldehyde-3-Phos-phate Dehydrogenase (gpd) and Beta-Tubulin (TUB)) toassess Guyanagaster and Armillaria systematic relation-ships through phylogenetic analysis. Not all loci wereavailable for all specimens. DNA sequences that werenot generated in this study were obtained from Gen-Bank or published whole genome sequences. Collectioninformation for all specimens and GenBank accessionnumbers for the included sequences are compiled inAdditional file 1.

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 3 of 16

Sequences were aligned in Mega 5.0 [37] using theMUSCLE algorithm [38] with refinements to the align-ment done manually. The introns of EF1α were alignableacross the Armillaria/Guyanagaster dataset and weretherefore included. Phylogenies were reconstructedusing maximum likelihood (ML) and Bayesian methods.The GTR +G model of molecular evolution was selectedfor all data sets as determined by PartitionFinder v1.1.0[39]. Maximum likelihood bootstrap analysis for phyl-ogeny and assessment of the branch support by boot-strap percentages (BS%) was performed using RAxMLv2.2.3 [40]. One thousand bootstrap replicates wereproduced. Each locus was analyzed separately as well asall together in a supermatrix data set using ML. Bayesiananalyses for the reporting of Bayesian posterior prob-ability (BPP) support for branches was conducted onindividual loci as well as the six-gene data set usingthe program Mr. Bayes v3.2.2 [41]. Four simultan-eous, independent runs each with four Markov chainMonte Carlo (MCMC) chains, were initiated and runat a temperature of 0.15 for 20 million generations,sampling trees every 1000 generations until the stand-ard deviation of the split frequencies reached a finalstop value of 0.01. We discarded the initial 10% oftrees as burn in and produced a maximum cladecredibility tree from the remaining trees; 0.95 BPPrepresents a well-supported lineage.

Divergence time estimationDivergence ages within the armillarioid clade were esti-mated using the fossil calibration approach describedand implemented by [24, 42–44]. Molecular clock ana-lysis was performed using BEAST v1.8.3 [45] with XMLfiles containing BEAST commands and priors assembledin BEAUTi v1.8.3. These files included the followinganalytical settings: GTR +G model of evolution, uncor-related relaxed clock with lognormal rate distribution;tree prior was set to speciation birth-death process, run-ning 100 million generations, sampling every 1000th tree.The analysis was run three times. The first 10% of treeswere removed as burn-in after ensuring a minimumeffective sample size (ESS) of 200 was reached for all ofthe parameters. The remaining trees—representing theposterior distribution from all Bayesian analyses—werecombined into a single file using LogCombiner v1.8.3.This file was used to produce a summary tree using TreeAnnotator v1.8.3. The mean ages of supported nodes, aswell as the corresponding 95% highest posterior dens-ities (HPDs), were examined from BEAST logfiles usingTracer v1.6 [46].Taxa used in this analysis are in bold in Additional file

1. It includes 21 Armillaria species (one representativespecimen from each of the species in the first analysis),two Guyanagaster specimens (one representative from

each collecting locality), nine Physalacriaceae species(one representative from nine different genera that en-compass the major lineages within this family fide [47]),while six species from Agaricales families closely relatedto the Physalacriaceae fide [48] served as outgroup taxa.The alignment was analyzed using five partitions: 28Swas analyzed in a single partition, while two codon parti-tions ((1 + 2), 3) were utilized for both EF1α and RPB2.The marasmioid fungi (Marasmius alliaceus, Marasmiusrotula and Mycena amabilissima) were calibrated basedon a 90-Ma fossil Archaeomarasmius leggetti from mid-Cretaceous amber [49], following the parameter settingsof [24]. In BEAUTi, this prior was set as a lognormaldistribution with a mean of 10, log standard deviation of1 and offset of 90, with mean in real space, truncated onthe lower end at 90 and at the upper end at 200.

Ancestral range estimationTo estimate the ancestral range of the armillarioid cladeand lineages within, we used dispersal-extinction-cladogenesis (DEC) analysis in the package Lagrange[50] and implemented in the program RASP [51]. Sixareas were defined: North America, Europe + Asia (here-inafter referred to as Eurasia), Africa, tropical SouthAmerica (composed of the area covered by tropical rain-forest in South America), Australasia (composed ofAustralia, New Zealand and southeast Asia) and temper-ate South America (composed of the area not coveredby tropical rainforest in South America). Each taxon inour data set was assigned to areas based on its currentknown range. In the scenario we tested, movement be-tween any area at any time was unconstrained. Underthis model, all areas are treated as equally probable an-cestral ranges. This model was tested under both twoand three area constraints. The consensus tree producedduring the divergence time analysis was used for allLagrange calculations and the root node of the armillar-ioid clade was calibrated to 51 Ma.

ResultsSequences, collections and culturesEight ITS, eleven 28S, twelve EF1α and five RPB2 se-quences were generated during this study (Additionalfile 1). The size of the sequences ranged from 514–713,705–1031, 796–1233, and 553–704 bp, respectively.After the ends of the individual alignments weretrimmed, the size of the aligned datasets were as follows:28S was 906 bp; EF1α was 906 bp; RPB2 was 638 bp;actin-1 was 607 bp; gpd was 499 bp; TUB was 911 bp.The six-gene dataset was composed of a total of 52Armillaria specimens, representing 21 species, as well asfour Guyanagaster specimens, representing two species.The number of taxa in each of the single-locus phyloge-nies is as follows: 28S had 42 taxa; EF1α had 57 taxa;

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 4 of 16

RPB2 had 32 taxa; actin-1 had19 taxa; gpd had 19 taxa;TUB had 12 taxa. Phylogenies for each locus, as well asa combination of loci, are available in Additional file 2.Two Armillaria specimens were collected from Guya-

na—MCA 3111 and TH 9751. ITS sequences from bothshared 99% identity with Armillaria puiggarii (GenBank:FJ664608). No significant morphological differenceswere found when comparing our field descriptions tothe species description by [52]. Two Armillaria speci-mens were collected from Cameroon—TH 9926 and THDJA 91. ITS sequences from both shared 99% identitywith an unidentified Armillaria species collected fromZimbabwe (GenBank: AY882982). Additionally, therewas variation between the two sequences. No significantmorphological differences were found when comparing TH9926 to the species description of Armillaria camerunensis[53, 54], but there was some morphological variation whencomparing TH DJA 91. For now, we conservativelyhypothesize that both of these specimens represent Armil-laria camerunensis. Morphological descriptions of TH9926 and TH DJA 91 are available in Additional file 3.Pure vegetative cultures were obtained from Guyanaga-

ster specimens MCA 3950 and RAK 88, as well as Armil-laria puiggarii specimen TH 9751, with rhizomorphsproduced in each (Fig. 1). After one month, differences ingrowth and branching pattern of the rhizomorphs can beobserved between the cultures of the two Guyanagasterspecimens, but the rhizomorphs of neither were observedto become melanized. In contrast, the rhizomorphs in theculture of A. puiggarii became melanized almost com-pletely. Strain MCA 3950 is deposited in Centraalbureauvoor Schimmelcultures with the strain accession numberCBS 138623, while strains RAK 88 and TH 9751 are avail-able from the authors upon request.

Resolved Guyanagaster phylogenetic hypothesisNo single locus was sufficient to obtain a well-supportedarmillarioid phylogeny (see Additional file 2), but throughthe use of multiple loci, we were able to obtain a well-supported armillarioid phylogeny (Fig. 2). The six-gene

analyses recovered a strongly supported armillarioid clade(100% BS and 1.00 BPP) (Fig. 2), which is composed of allknown Armillaria and Guyanagaster species. The armil-larioid clade is further composed of three well-supportedlineages that were recovered from both analyses (Fig. 2):(1) the annulate lineage (100% BS and 1.00 BPP), whichincludes annulate mushroom-forming armillarioid species,(2) the exannulate lineage (91% BS and 1.00 BPP), com-posed of exannulate mushroom-forming armillarioid spe-cies formerly placed in Armillaria subgen. Desarmillariaand (3) the gasteroid lineage (100% BS and 1.00 BPP),composed of Guyanagaster species. The gasteroid lineageconsists of two distinct species: G. necrorhizus from thePotaro region and an undescribed Guyanagaster speciesfrom the Mabura region of Guyana. These two species aregeographically separated by mountains of elevations over1000 m, as well as Kaieteur Falls and the Essequibo River,all within a distance of 125 km.The annulate lineage is further composed of four well-

supported lineages (Fig. 2), with species compositions asfollows: (1) African lineage composed of Armillaria fus-cipes and A. camerunensis; (2) melleioid lineage com-posed of A. mellea; (3) northern hemisphere lineagecomposed of A. altimontana, A. borealis, A. calvescens,A. cepistipes, A. gallica, A. gemina, A. sinapina and A.solidipes; and (4) Australasian + temperate South Ameri-can lineage composed of A. affinis, A. fumosa, A. hinnu-lea, A. limonea, A. luteobubalina, A. novae-zelandiae, A.pallidula and A. puiggarii.The most glaring conflict between single gene and

multi-gene analyses is the placement of the Africanlineage: in gpd, actin-1 and EF1α this lineage is nestedwithin or sister to the Australasian + temperate SouthAmerican lineage; in TUB and RPB2, it is sister to A.mellea; in 28S it is sister to Guyanagaster. In the multi-gene analyses, it is consistently recovered as divergingearlier than the A. mellea lineage. In all of the singlegene analyses except for 28S, a close relationshipbetween the exannulate and the gasteroid lineages wasrecovered.

Fig. 1 a Vegetative culture of Guyanagaster necrorhizus MCA 3950 showing unmelanized rhizomorphs. b Vegetative culture of Guyanagaster lucianiiRAK 88 showing unmelanized rhizomorphs. c Vegetative culture of Armillaria puiggarii TH 9751 showing melanized rhizomorphs. Bar = one cm

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 5 of 16

Age and ancestral range of Armillaria and GuyanagasterTable 1 has a complete list of the divergence dates forthe well-supported nodes recovered in the phylogen-etic analysis. Figure 3 is the time-calibrated phylogenyfor the group. The time to most recent common an-cestor (tMRCA) of the armillarioid clade estimated inthe BEAST analysis was 51 Ma (node 1; 95% HPD

30–73 Ma). The tMRCA of the exannulate lineagewas estimated at 41 Ma (node 4; 95% HPD 24–59 Ma).The tMRCA of the annulate armillarioid lineage wasestimated at 33 Ma (node 5; 95% HPD 19–47 Ma).The date of the divergence between the two Guyana-gaster species was estimated at 8 Ma (node 2; 95%HPD 3–14 Ma).

a

b

d

e

f

h i j

g

c

Fig. 2 Phylogram generated from the analysis of six gene regions (28S, EF1α, RPB2, TUB, gpd and actin-1) from 58 taxa. Guyanagaster is the earliestdiverging lineage and is sister to the mushroom-forming armillarioid species. The exannulate armillarioid species are the next to diverge, and composeDesarmillaria. The annulate armillarioid species form a monophyletic lineage and compose Armillaria sensu stricto. The major lineages within Armillaria areindicated in bold text at the corresponding node. The exannulate armillarioid species are sister to the annulate armillarioid species. Strobilurus esculentusand Oudemansiella mucida were selected as outgroup taxa. Black circles represent support of 90% (maximum likelihood bootstrap values, shown aspercentages) and 0.95 BPP (Bayesian posterior probabilities) or greater, grey circles represent support of 0.95 BPP or greater and white circles representsupport of 75% or greater. Images: (a) Armillaria sinapina; (b) Armillaria hinnulea; (c) Armillaria puiggarii; (d) Armillaria mellea; (e) Armillaria camerunensis; (f)Desarmillaria tabescens; (g) Desarmillaria ectypa, (h–j) Guyanagaster necrorhizus. (Photo credits: (a) Christian Schwarz; (b) J. J. Harrison; (c) Todd F. Elliott; (d,h–j) Rachel A. Koch; (e) Terry W. Henkel; (f) Stephen D. Russell; (g) Tatyana Svetasheva)

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 6 of 16

Table 1 Ages for nodes, confidence intervals and posterior probability on the phylogenetic tree presented in Fig. 3

Node Lineage Mean Age 95% HPD [min, max] Posterior Probability

1 Armillarioid 50.81 30.03, 72.32 1.00

2 Guyanagaster 7.88 2.70, 13.91 1.00

3 Desarmillaria + Armillaria 41.32 24.47, 59.07 0.90

4 Desarmillaria 31.74 15.83, 47.71 1.00

5 Armillaria 33.00 18.90, 46.98 1.00

6 African lineage 15.49 6.84, 24.76 1.00

7 Melleioid lineage 30.12 17.73, 43.43 0.90

8 Northern hemisphere + Australasian + temperate South American lineage 27.52 17.13, 37.60 0.94

9 Northern hemisphere 12.07 6.12, 18.87 1.00

10 A. borealis + A. solidipes + A. gemina 6.66 2.41, 11.51 1.00

11 A. altimontana + A. calvescens + A. gallica 8.58 3.80, 14.02 1.00

12 A. cepistipes + A. sinapina 6.44 2.22, 11.24 0.52

13 Australasian + temperate South American 19.27 11.18, 27.99 0.98

14 A. novae-zelandiae + A. affinis + A. puiggarii 8.96 4.04, 14.47 1.00

15 A. hinnulea + A. pallidula + A. fumosa + A. limonea + A. luteobubalina 18.13 10.30, 26.29 –

Fig. 3 Time-calibrated phylogeny generated from Bayesian analysis of three gene regions (28S, EF1α, RPB2) from 19 Armillaria species, twoDesarmillaria species and two Guyanagaster species. Geographic origin of each specimen is indicated by the box to the left of the name:red = Neotropics, blue = Australasia, brown = temperate South America, lime green = Eurasia, purple = North America, and turquoise = Africa.Boxes at ancestral nodes correspond to the most probable ancestral range at that node as presented in Additional file 4. Geologic epochs are notedabove the time scale. Numbers at nodes correspond to lineages in Table 1 and Additional file 4. Dark grey bars correspond to the 95% HPD andcorrespond to the values in Table 1. The map in the lower left hand corner represents the proposed dispersal pattern

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 7 of 16

We estimated the ancestral ranges for 15 well-supported clades. The most probable area was the sameunder range constraints of ≤ 2 and ≤ 3 for all nodes. Themost probable ancestral range for the most recent com-mon ancestor (MRCA) of the armillarioid clade is Eurasia.Eurasia is also the most probable ancestral range for all ofthe annulate mushroom-forming lineages containingnorthern hemisphere taxa. Migration to both Africa andAustralasia from Eurasia is the most probable scenario asestimated from this analysis. Additional file 4 provides theprobabilities for each area being the ancestral range underboth area constraints; numbers in bold represent the mostprobable region at that node.

DiscussionOur phylogenetic analyses, coupled with morphologicaldata, support a three-genus composition of the armillar-ioid clade: Armillaria contains the annulate mushroom-forming species, Desarmillaria contains the exannulatemushroom-forming species, and Guyanagaster containsthe known gasteroid species. Our estimates of thetMRCA of the armillarioid clade suggest it arose 51 Ma,which is in accord with a previous estimate [10], eventhough a different set of taxa, loci and calibrationmethods were used. During this time, Earth was experi-encing one of the warmest intervals of the past 65 mil-lion years known as the Paleocene-Eocene ThermalMaximum (PETM) [55]. This warm and humid climateled to the widespread expansion of area covered in trop-ical rainforests. The ancestral range estimation suggeststhe most probable area for the origin of the MRCA isthe Eurasian subcontinent (Additional file 4), much ofwhich contained tropical rainforests at that time [56].

Biogeography and evolution of Guyanagaster andDesarmillariaOur phylogenetic reconstruction suggests that Guyanaga-ster species evolved within the earliest diverging lineage inthe armillarioid clade. The process of gasteromycetation isbelieved to be unidirectional; gasteroid fungi evolve from aballistosporic ancestor, and there is no evidence that thistransformation has ever been reversed [11]. With this inmind, the ancestor to the armillarioid lineage was likely agilled mushroom. According to our estimations, between51 and 8 Ma the mushroom progenitor to Guyanagasterunderwent the gasteromycetation process, resulting in theevolution of Guyanagaster. At one point in time, thisancient lineage was composed of mushroom-forming spe-cies—all of which likely went extinct—as the only knownextant members of this lineage are Guyanagaster species.The next lineage to diverge is composed of the two

described Desarmillaria species: D. ectypa and D.tabescens. Neither of the species in this genus havean annulus at maturity, in contrast to all other known

mushroom-forming armillarioid species. Desarmillariaectypa is different from all other armillarioid speciesby virtue of its ecology: whereas Armillaria speciesare pathogens and wood decomposers, D. ectypa isrestricted to peat bogs and associated with Sphagnum[57]. The ecology of D. tabescens is much more typ-ical of Armillaria species. It is a primary pathogentowards introduced Eucalyptus species in France, aswell as a secondary pathogen to Quercus species, butin many cases it is a saprotroph, colonizing stumps ofQuercus species [58].Both Guyanagaster and Desarmillaria evolved within

the oldest lineages in the armillarioid clade and bothgenera contain two known species. Armillaria repre-sents the most recently diverged genus-level lineagewithin the armillarioid clade, yet has over 35 describedspecies (see [59] for species counts). The older ages ofthe lineages from which Guyanagaster and Desarmil-laria evolved, coupled with the fact that they containfar fewer extant species, suggest these two genera areon a different evolutionary trajectory compared toArmillaria. As stated above, our ancestral range estima-tion suggests the most probable ancestral range for theMRCA to the armillarioid clade is Eurasia, whereas spe-cies of Guyanagaster have only so far been found inGuyana. One question is how the mushroom progeni-tor to Guyanagaster dispersed from Eurasia to Guyana.One possibility is that the mushroom progenitor toGuyanagaster was widely dispersed within the humid,tropical rainforest habitat that predominated at thetime it arose. It is possible that Guyanagaster evolvedwithin the Guiana Shield and represents a highly spe-cialized endemic to this region.The two species of Desarmillaria are thought to be

more thermophilic than their sympatric Armillariaspecies [57, 60]. This temperature adaptation could bea residual characteristic of their origin during thePETM and could also explain the paucity of speciesin this genus. As the climate changed, species adaptedfor a cool and arid climate could have competitivelydisplaced the species adapted for a warm and humidclimate, in this case, species of Desarmillaria and themushroom progenitor to Guyanagaster. The novelmorphology of Guyanagaster species suggests thatthey are highly specialized for their current habitat,and as stated above, the restriction of D. ectypa topeat bogs is a specialized habitat for armillarioidfungi. Habitat specialization has been shown to act asbuffer from extinction [61], and could be what hasallowed species in these early diverging lineages toavoid competitive displacement by better-adaptedarmillarioid fungi. Whether D. tabescens occupies aspecialized niche that has aided in its persistence tothe present remains unknown.

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 8 of 16

Biogeography and spread of ArmillariaAccording to our analyses, Armillaria sensu strictostarted diversifying approximately 33 Ma in Eurasia, atwhich time the Earth’s climate shifted from the relativelyice-free world to one with glacial conditions; the tropicalclimate during which the MRCA of the armillarioidclade arose was replaced by a period of severe cooling anddrying as well as an increase in seasonality [55]. This ledto a contraction in the area covered by tropical rainforests[62] and the geographic expansion of species-poor decidu-ous vegetation [63]. Therefore, we hypothesize that Armil-laria species represent annulate mushrooms thatsurvived, diversified and radiated with their hosts as theyadapted to the drier, cooler and seasonal climate (i.e. tem-perate) of the late Eocene and beyond.Armillaria species in Africa form a monophyletic

lineage (100% and 1.00 BPP). Single gene analyses andmulti-gene analyses were ambiguous as to where thislineage diverged within the armillarioid clade (Additionalfile 2). In all multi-gene combinations, the AfricanArmillaria lineage is resolved as the earliest divergingwithin Armillaria sensu stricto, which is similar to previ-ous work that has also suggested that this lineage isdivergent compared to the rest of the armillarioidlineage [10, 64]. Our ancestral area estimation suggeststhat the ancestral region of the African lineage is Eur-asia. During the Eocene and Oligocene, a shallow seawayseparated Africa from Eurasia, acting as a barrier to dis-persal. In the early Miocene (19–12 Ma), the seawaydrained, effectively linking the biotas in these tworegions [65]. Our molecular clock analysis suggests thatextant species in Africa started diversifying approxi-mately 15 Ma, which fits a migration from Eurasiaduring this time.Many northern hemisphere Armillaria species are

found in both North America and Eurasia (see [58, 59]).In a study of A. mellea, which is a species distributedthroughout the northern hemisphere, four distinct line-ages were identified, corresponding to their locality: 1)western North America, 2) eastern North America, 3)western Eurasia and 4) eastern Eurasia [66]. They sug-gest that this species was once widespread and is now inthe process of speciation [66]. Within this backdrop, it ispossible that Armillaria species facilitated their dispersalas pathogens on the diverse flora that existed during thistime. As available pathways disappeared, ultimately ceas-ing migration, the species started to diverge. From ourancestral range estimation, we can see that there weremultiple armillarioid introductions into North America(Fig. 3), suggesting a dynamic process of dispersal acrossthe northern hemisphere. A more comprehensive studyon northern hemisphere Armillaria species is necessaryto determine when and how they migrated between themajor landmasses.

Armillaria species started diversifying in the temperatesouthern hemisphere (outside of Africa) about 19 Ma(Fig. 3, node 13). Our results suggest that the most prob-able dispersal pathway was from Eurasia to Australasia.In a plant biogeographic meta-analysis of the Australa-sian region, it was found that after 25 Ma, there was anincrease of flora inputs to Australia from the IndoMala-yan region, representing the time when the continentalshelf containing Southeast Asia collided with that con-taining New Guinea and Australia [67]. The authors ofthis study hypothesized that habitat instability in the re-gion at that time generated vacant niches and increasedthe probability of successful establishment of dispersedlineages compared with the relatively stable, saturatedstates existing prior to approximately 25 Ma [67]. In thisscenario both the greater dispersal distance and relativedifficulty in establishing viable populations together con-tributed to the infrequency of successful exchanges priorto 25 Ma. This too follows our hypothesis that Armil-laria species dispersed as pathogens on flora that existedduring this time.Many of the same Armillaria species are found in both

Australasia and temperate South America [68]. This pat-tern fits with an ancient Gondwanan distribution, butthe lineage is far too young for this to be the case. Alter-natively, the discovery of fossilized Nothofagus wood inAntarctic deposits from the Pliocene [69] suggests thiscould have been an overland dispersal route until 2 Ma.Many of the fungi in this Australasian + temperate SouthAmerican lineage are thought to be close associates withNothofagus [68], suggesting again that they could havedispersed with their plant associates through Antarctica.Alternatively, long distance basidiospore dispersal eventscould account for the geographic disjunction in closelyrelated Armillaria species. Because appropriate hos-ts—species of Nothofagus—occur in both Australasia andtemperate South America [70], establishment after a long-distance basidiospore dispersal event could be moreprobable.

The rise of rhizomorphs: morphological developments inthe armillarioid cladeOur phylogenetic hypothesis sheds light on the evolutionof rhizomorphs in the armillarioid clade. Rhizomorphs arediscrete, filamentous aggregations that extend from a re-source base into substrates that may not support theirgrowth, foraging for new resource bases [71]. Individualsthat produce rhizomorphs can occupy huge swaths ofhabitat and prevent other organisms from establishing, asis the case with the “humungous fungus” [72]. All armil-larioid species have the capacity to form rhizomorphs inculture, but only the later diverging lineages have been ob-served to form them in nature (see Table 2, Fig. 1).

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 9 of 16

Table 2 Trophic strategy, rhizomorph production in nature and known geographic range of species used in this study

Species FacultativeNecrotroph?

Characterization Rhizomorphsin nature?

Hosts Known range References

Armillariaaffinis

Unknown Unknown Central America [59]

Armillariaaltimontana

Unknown Yes Hardwoods andconifers

Higher elevation forestsof Western interior ofNorth America

[83]

Armillariaborealis

Yes Weakly pathogenic,opportunistic

Yes Birch, wild cherry Europe [58, 84]

Armillariacalvescens

Yes Opportunistic pathogenon stressed trees

Yes Hardwoods, particularlysugar maple

Eastern North America [85, 86]

Armillariacamerunensis

Unknown Observed in a disease center,but unknown if it is the causalagent

None observed Unknown Africa This study

Armillariacepistipes

Yes Weakly pathogenic Yes Conifers Europe, North America [58, 87]

Armillariafumosa

Unknown Unknown Australia [59]

Armillariafuscipes

Yes Particularly pathogenicto exotic species

None observed Pine forest plantations,Acacia and Cordia species

Africa, India [74, 75, 88]

Armillariagallica

Yes Weakly or secondarilypathogenic

Yes Hardwoods Europe, North America,Asia

[58]

Armillariagemina

Yes Primary pathogen Yes Maples, beech Birch Eastern North America [86]

Armillariahinnulea

Yes Secondary pathogen Yes In wet sclerophyll forests Australia, New Zealand [89, 90]

Armillarialimonea

Yes Pathogenic to pine seedlings(introduced tree)

Yes Pine Argentina, Chile,New Zealand

[68, 91, 92]

Armillarialuteobubalina

Yes Primary pathogen in nativeforests

Yes Eucalyptus Australia, Tasmania [68, 93–95]

Armillariamellea

Yes Highly pathogenic Yes Over 600 ornamentals,hardwood and orchardtrees

Europe, North America,Asia

[58, 96]

Armillarianovae-zelandiae

Yes Pathogenic to pine seedlings(introduced tree)

Yes Pine Argentina,Australia, Chile, NewZealand

[91, 92, 95]

Armillariapallidula

Unknown Unknown Australia [59]

Armillariapuiggarii

Unknown Observed in a disease center,but unknown if it is the causalagent

Melanizedrhizomorphsobserved in the field

Dicymbe spp. Argentina, BoliviaCaribbean, Guyana

[95] andthis study

Armillariasinapina

Yes Weakly pathogenic Yes Conifers North America, Japan [97]

Armillariasolidipes

Yes Highly pathogenic Yes Conifers Cooler regions ofNorth America, Europe,China

[58, 98]

Desarmillariaectypa

Unknown Saprotrophic on decayingpeat moss

No Sphagnum moss Europe, Russia, Japan,China

[57]

Desarmillariatabescens

Yes Highly pathogenic No Eucalyptus, Quercus Asia, Europe,North America

[58, 73]

Guyanagasterlucianii

Unknown Only saprotrophic stageobserved

No Eperua spp. Guyana This study

Guyanagasternecrorhizus

Unknown Only saprotrophic stageobserved

No, but short non-melanized hyphalcords may beproduced

Dicymbe spp. Guyana [27]

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 10 of 16

Species in the earliest diverging genera in the armillar-ioid clade, Desarmillaria and Guyanagaster, have notbeen observed to form rhizomorphs in nature, but theydo produce unmelanized rhizomorphs in culture [27, 57,58, 73] (Fig. 1). Rhizomorph production of Armillariaspecies from Africa, the next lineage to diverge, has beenscantily observed in nature [74], but the production ofmelanized rhizomorphs does occur in culture [75]. Thenext lineage to diverge is that composed of A. mellea,which produces short-lived melanized rhizomorphs withlimited growth in the soil [58]. The most species rich armil-larioid lineage, the northern hemisphere and Australasian/temperate South America lineage, produce melanized rhi-zomorphs in nature (see Table 1).The capacity of Armillaria species to produce mela-

nized rhizomorphs in nature is thought to be an adapta-tion to harsh environments [71]. Melanized rhizomorphscan confer advantages like protection against other mi-crobial competitors, translocation of resources, growthfrom a suitable resource base into an environment thatmay not support growth, as well as enhancement of in-oculum potential [71]. Additionally, melanin is thoughtto promote longevity and survival of rhizomorphs withinthe soil [76], and has also been found to help other fungisurvive in extreme environments [77]. Although allspecies in the armillarioid clade appear to retain the traitof rhizomorph production, only the most recentlydiverged appear to be adapted for melanized rhizomorphproduction in the current environment. In a controlled en-vironment, it was found that only under high oxygen avail-ability and near saturated moisture would D. tabescensproduce melanized rhizomorphs [78]. It was concluded thatthese environmental conditions are sufficiently stringent inthe climate of today, and could explain why D. tabescenshas not been observed to form rhizomorphs in nature. It ispossible that the production of melanized rhizomorphscould be a driver of diversification in Armillaria by confer-ring a competitive advantage to the species that producethem.Desarmillaria species lack an annulus at maturity,

whereas Armillaria species have a robust annulus at ma-turity. An annulus is the remnant of a partial veil thatremains after the pileus expands. Partial veils protect theimmature hymenium [79]. The development of a moreprotected hymenium could be an adaptation of Armil-laria to drier, more unpredictable habitats that werecommon when this lineage began to diversify.

ConclusionsOur analyses suggest that the armillarioid clade arose inEurasia during the PETM, a time marked by a warm,tropical climate. Two lineages arose during this time: theearliest diverging lineage—which eventually led to thegasteroid genus Guyanagaster—and the exannulate

mushroom-forming genus Desarmillaria. Besides beingthe oldest lineages in the armillarioid clade, they are alsodepauperate compared to Armillaria. Armillaria divergedafter the shift to a cooler and more arid climate at theEocene-Oligocene boundary. The production of mela-nized rhizomorphs in nature and the development of aprotective partial veil could be adaptations that led to thesubsequent dispersal and diversification of Armillaria inthe much harsher, temperate climate. The success ofArmillaria could have displaced now-extinct species inthe exannulate and gasteroid lineages. Guyanagaster andDesarmillaria species have likely persisted to the presentbecause they are highly specialized for their habitat.

Formal taxonomic descriptionsDesarmillaria (Herink) R. A. Koch & Aime gen. etstat. nov.Basionym—Armillaria ss. Fries subgenus Desarmillaria

Herink, Sympozium o václavce obecné (J. Hasek). 1972September. Lesnicka fakulta VSZ Brno: 44. 1973.Type—Armillaria socialis (DC. ex Fr.) Herink, Sympo-

zium o václavce obecné (J. Hasek). 1972 September.Lesnicka fakulta VSZ Brno: 44. 1973.MycoBank number—MB 819124 (genus).Description—Basidiomata stipitate, usually in caespi-

tose clusters on wood. Pileus convex to applanate, usu-ally some shade of brown. Lamellae pallid, adnexed toadnate to subdecurrent with lamellulae. Stipe central,various colorations but often concolorous with cap, lon-gitudinally striate, usually with fibrils on upper third.Annulus absent. Basidiospores ellipsoid to spherical,smooth, hyaline, inamyloid. Basidia clavate, four-sterigmate, hyaline. Cheilocystidia clavate, usually resem-bling basidioles. Pileipellis suprapellis composed ofround, ellipsoid, cylindric, utriform, brown to hyaline,verrucose cells; subpellis composed of a layer ofcompact, shortened hyphae. Clamp connections absent.Rhizomorph production not observed in nature, whileunmelanized production in culture has been observed.Saprotrophic to parasitic. Known only from the northernhemisphere.Commentary—Desarmillaria includes mushroom-

forming armillarioid species that lack an annulus. Thisdifference in morphology led Singer [80, 81] to divideArmillaria (as Armillariella) into two sections based onthe presence or absence of an annulus at maturity.Herink [82], then, recognized Armillaria as an annulatesubgenus and Desarmillaria as an exannulate subgenus.Additionally, members of this genus have not been ob-served producing rhizomorphs in the field [57, 58], whichis in contrast to most Armillaria species, most of whichdo produce rhizomorphs in the field (see Table 2). Rhizo-morph production in nature appears to be a lost trait inthis genus, as species are still able to form them in culture.

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 11 of 16

We have opted to elevate Desarmillaria to genus-levelfor two reasons. First, the absence of an annulus inDesarmillaria species and the presence of one in Armil-laria species is a reliable characteristic to differentiatethe two genera. Second, our phylogenetic and molecularclock analyses show Desarmillaria is on a separate evo-lutionary trajectory compared to Armillaria, meriting aseparate genus.Desarmillaria tabescens (Scop.) R. A. Koch & Aime

comb. nov.Basionym—Agaricus tabescens Scop., Scopoli 1772, Flora

Carniolica Plantas Corniolae Indigenas no. 1537: 446.MycoBank number—MB 819125 (species).Desarmillaria ectypa (Fr.) R. A. Koch & Aime

comb. nov.Basionym—Agaricus ectypus Fr., Fries 1821, Syst.

Mycol. 1: 108.MycoBank number—MB 819126 (species).Guyanagaster lucianii R. A. Koch & Aime sp.

nov.—Holotype—BRG 41292, isotype—PUL F2891(Figs. 1 and 3b–h). Mabura Ecological Reserve, (5°10'09.26"N, 58°42'14.40"W), 25 December 2013. R. A.Koch 89.MycoBank number—MB 815807 (species).Representative DNA barcode—RAK 89 (holotype),

ITS—GenBank KU170950, 28S—GenBank KU170940,EF1α—GenBank KU289110.Etymology—Lucianii = in honor of “Lucian” Edmund, a

Patamona fungal parataxonomist, who discovered thetype locality of G. lucianii and has since been invaluablein collecting Guyanagaster specimens in the field.

Description—Figures 2 and 4a–g. Basidiomata gaster-oid, subhypogeous to hypogeous, scattered, or in lineartroops, attached directly to woody roots of Eperuafalcata and Dicymbe altsonii trees; 12–34 mm broad, 8–25 mm tall, globose to subglobose to ovoid and irregu-larly broadly lobate, dense, base with smooth, sterileconcolorous stipe, 5–44 × 2–5 mm, attached directly tothe substratum. Peridium dark brown (8 F4-8 F3) toblack (8 F1) during all stages of development, moist,tough, covered in 4–6 sided polygonal pyramidal wartsthat come to a shallow point, these 0.8–3.1 mm broad,0.5 mm tall, sharply differentiated from endoperidium;endoperidium white (8A1) to light pink (8A2), tough,0.5–1 mm thick, composed of matted hyphae transition-ing evenly to sterile, white hyphal veins between the gle-bal locules; gleba composed of well-defined locules andintervening veins of hyphae; locules globose to ovate tosubangular, 0.2–2 mm broad, initially white (8A1), ma-turing in stages to gold (3A2-4A5), light orange-pink(7B6), to deep orange (8B8), often evenly, but rarely thelocules that border the peridium and columella take onthe darker shade first; hyphae separating locules initiallywhite (8A1), hyaline and waterlogged in mature speci-mens; columella well-defined to not, base continuouswith stipe, 6–13 × 3–6 mm, initially white (8A1), turningto brown (6D7), taking on the color of the gleba withage, cottony to gelatinous. Exoperidium individualhyphae of outer layer dark brown, individual hyphaecurved to tortuously curved, thick-walled, non-gelatinous,18–99 × 4–11 μm. Endoperidium 700–800 μm thick, well-differentiated from exoperidium and glebal locules, hyphae

Fig. 4 a–g Guyanagaster lucianii (BRG 41292 HOLOTYPE). a Exterior showing dark brown peridium and irregularly lobed shape. b Subhypogeoushabit. c Longitudinal section of basidioma showing immature gleba, developed columella and elongated stipe. d Maturing gleba. e Basidiosporesof immature basidioma. f Basidiospores of mature basidioma. g Mature gleba. Bar = one cm in a, b, c, d and g. Bar = 20 μm in e and f

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 12 of 16

hyaline, 3–11 μm. Sterile hyphae separating locules similarto endoperidial hyphae, hyaline, thin-walled, branching, 3–5 um wide, organized in a parallel manner. Basidia notobserved. Basidiospores globose, dextrinoid; 21.4–30.4 ×21.4–29.3 μm (mean = 25.1 ± 1.5 × 24.8 ± 1.5 μm, Qr =0.95–1.10, Qm = 1.01 ± 0.04; n = 20); immature basidio-spores hyaline when immature, then darkening to light pinkand then rusty brown with maturity. Columella hyphaehyaline, thin-walled, not as tightly packed, some randomlyinflated at septa, 4–13 μm wide, random apical cellsinflated, 13–26 × 37–57 μm. Clamp connections absent inall tissues. Taste not tested. Odor not tested.Habit, habitat and distribution—Attached to decaying

woody roots of Dicymbe altsonii and Eperua falcata intropical rainforests. Known only from the type locality inthe Mabura Ecological Reserve of Guyana.Specimens examined—Guyana. Region 10 Upper

Demerara-Berbice—Mabura Hill, elevation 161 m; vicinityof basecamp, Eduba stand 4, four basidiomata, 11 June2011, MCA 4424 (BRG 41294, PUL F3439); vicinity ofbasecamp, patch 32, one basidioma found at the base of adead Eperua falcata, 24 December 2013, RAK 84 (BRG41230, PUL F3440); vicinity of basecamp, patch 35, fourbasidiomata found at the base of a dead Eperua falcata,24 December 2013, RAK 88 (BRG 41293, PUL F2892);2.5 km north of basecamp, patch 31, nine basidiomatafound along the decaying roots of Eperua falcata, 25December 2013, RAK 89 (BRG 41292, PUL F2891).Commentary— Phylogenetic evidence suggests that

the two lineages of Guyanagaster represent distinct spe-cies that diverged approximately 8 Ma. The sister speciesshare 87% (777/893 bp) nucleotide identity at the ITSregion and 96% (705/731 bp) nucleotide identity at LSU.The drastic morphological changes that occur duringthe maturation of Guyanagaster basidiomata (i.e. glebacoloration) as well as the morphological plasticity betweenindividual basidiomata (e.g. basidioma size and shape,columella dimensions) make delimiting species usingmacromorphological characters unreliable. However, G.lucianii can be distinguished from G. necrorhizus by itsconsistently larger basidiospores. During all stages of mat-uration, the basidiospores of G. lucianii measure between21–30 × 21–29 μm, compared to those of G. necrorhizus,which measure 15–19 × 15–18 μm. Additionally, the ba-sidiospores of G. lucianii lack the pedicel that is readily ap-parent on the basidiospores of G. necrorhizus (Fig. 4e–f ).Ecologically, G. lucianii appears to occupy the same

niche as G. necrorhizus. Basidiomata are hypogeous tosubhypogeous and growing from the roots of dead anddecaying trees. Both Guyanagaster species are white rot-ters as the decaying roots to which they are attached areall light-colored, spongy and sometimes gelatinous. Theknown habitat for G. necrorhizus and G. lucianii are sep-arated by only 125 km, so elucidation of their dispersal

strategy is needed to understand the evolutionary forcesthat led to their speciation.

Additional files

Additional file 1: Collection data for the specimens used in thephylogenetic analyses and GenBank accession numbers for thecorresponding sequence data, both generated for this study andobtained from previous studies [10, 64, 83, 99–114]. (XLSX 54 kb)

Additional file 2: Single-locus and multi-gene phylogenies for the sixloci used. (PDF 3262 kb)

Additional file 3: Morphological description of Armillaria camerunensisTH DJA 91 and TH 9926. (DOCX 108 kb)

Additional file 4: Results from the ancestral range estimation analysis.(XLSX 38 kb)

AcknowledgementsThe authors wish to thank the following for assisting with this study: DillonHusbands functioned as the Guyanese local counterpart and assisted with fieldcollecting, descriptions and specimen processing. Todd F. Elliott, J. J. Harrison,Stephen D. Russell, Christian Schwarz and Tatyana Svetasheva providedphotographs for Fig. 2. Additional field assistance in Guyana was provided by C.Andrew, V. Joseph, P. Joseph, F. Edmond, L. Edmond and L. Williams. InCameroon, Dr. Jean Michel Onana, Head of The National Herbarium of Cameroon(Institute of Agricultural Research for Development, IRAD), provided muchlogistical assistance. The Conservator of the Dja Biosphere Reserve, Mr.Mengamenya Goue Achille, and his staff greatly assisted the fieldwork in the Dja.Field assistance in Cameroon was provided by Alamane Gabriel (a.k.a. Sikiro),Abate Jackson, Mamane Jean-Pierre, Mei Lin Chin, Todd Elliott, and CamilleTruong. Finally, the authors wish to thank Brandon Matheny and four anonymousreviewers for their insightful comments on an earlier version of this manuscript.Research permits were granted by the Guyana Environmental Protection Agencyand by the Cameroon Ministry of Research and Scientific Innovation. This paper isnumber 217 in the Smithsonian Institution’s Biological Diversity of the GuianaShield Program publication series.

FundingThe authors wish to thank the following for financial support: NationalScience Foundation DEB-0918591, NSF DEB-0732968, NSF DEB-1556412and Linnaean Society Systematics Research grant to MCA, the NationalGeographic Society’s Committee for Research and Exploration grant 9235–13to TWH, Explorer’s Club Exploration Fund, Mycological Society of America Fun-gal Forest Ecology grant, American Philosophical Society Lewis and Clark Fundfor Exploration and the T. Woods Thomas award to RAK.

Availability of data and materialsThe alignment and tree files are available in TreeBase and can be accessedat http://purl.org/phylo/treebase/phylows/study/TB2:S20469. Sequencesgenerated in this study were deposited in GenBank and accession numbersfor those sequences are available in Additional file 1.

Authors’ contributionsRAK, AWW and MCA conceived and developed the study. RAK, OS, TWH andMCA were involved in specimen collection and field descriptions. RAK performedall of the molecular work and RAK and AWW performed the analyses. RAK, AWW,and MCA wrote the paper. All authors read, advised on revisions and approvedthe final manuscript.

Competing interestsThe authors declare that they have no competing interests.

Author details1Department of Botany and Plant Pathology, Purdue University, WestLafayette, IN 47907, USA. 2Sam Mitchel Herbarium of Fungi, Denver BotanicGardens, Denver, CO 80206, USA. 3Institute of Agricultural Research forDevelopment (IRAD), National Herbarium of Cameroon (MINRESI), PO Box1601, Yaoundé, Cameroon. 4Department of Biological Sciences, HumboldtState University, Arcata, CA 95521, USA.

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 13 of 16

dx.doi.org/10.1186/s12862-017-0877-3dx.doi.org/10.1186/s12862-017-0877-3dx.doi.org/10.1186/s12862-017-0877-3dx.doi.org/10.1186/s12862-017-0877-3http://purl.org/phylo/treebase/phylows/study/TB2:S20469

Received: 8 June 2016 Accepted: 10 January 2017

References1. Wargo PM. Armillariella mellea and Agrilus bilineatus and mortality of

defoliated oak trees. Forest Sci. 1977;23:485–92.2. Kile GA. Armillaria root rot in eucalypt forests: aggravated endemic disease.

Pac Sci. 1983;37:457–64.3. Podger FD, Kile GA, Watling R, Fryer J. Spread and effects of Armillaria

luteobubalina sp. nov. in an Australian Eucalyptus regnans plantation. T BritMycol Soc. 1978;71:77–87.

4. Baumgartner K, Rizzo DM. Spread of Armillaria root disease in a Californiavineyard. Am J Enol Viticult. 2002;53:197–203.

5. Redfern DB, Filip GM. Inoculum and infection. In: Shaw CG, Kila GA, editors.Armillaria Root Disease. Agriculture Handbook 691. Washington DC: USDAForest Service; 1991. p. 48–61.

6. Kile GA, McDonald GI, Byler JW. Ecology and Disease in Natural Forests. In:Shaw CG, Kila GA, editors. Armillaria Root Disease. Agriculture Handbook691. Washington DC: USDA Forest Service; 1991. p. 102–21.

7. Moncalvo JM, Buchanan PK. Molecular evidence for long distance dispersalacross the Southern Hemisphere in the Ganoderma applanatum-australespecies complex (Basidiomycota). Mycol Res. 2008;112(4):425–36.

8. James TY, Moncalvo JM, Li S, Vilgalys R. Polymorphism at the ribosomalDNA spacers and its relation to breeding structure of the widespreadmushroom Schizophyllum commune. Genetics. 2001;157:149–61.

9. Geml J, Timling I, Robinson CH, Lennon N, Nusbaum HC, Brochmann C,et al. An arctic community of symbiotic fungi assembled by long-distancedispersers: phylogenetic diversity of ectomycorrhizal basidiomycetes inSvalbard based on soil and sporocarp DNA. J Biogeogr. 2012;39:74–88.

10. Coetzee MPA, Bloomer P, Wingfield MJ, Wingfield BD. Paleogene radiationof a plant pathogenic mushroom. PLoS One. 2011;6:e28545.

11. Hibbett DS, Pine EM, Langer E, Langer G, Donoghue MJ. Evolution of gilledmushrooms and puffballs inferred from ribosomal DNA sequences. ProcNatl Acad Sci U S A. 1997;94:12002–6.

12. Maser C, Maser Z, Molina R. Small-mammal mycophagy in rangelands ofCentral and Southeastern Oregon. J Range Manage. 1988;41:309–12.

13. Miller Jr OK, Miller HH. Gasteromycetes—Morphological and developmentalfeatures with keys to the orders, families, and genera. Eureka: Mad River Press;1988.

14. Bruns TD, Fogel R, White TJ, Palmer JD. Accelerated evolution of a false-trufflefrom a mushroom ancestor. Nature. 1989;339:140–2.

15. Mueller GM, Pine EM. DNA data provide evidence on the evolutionaryrelationships between mushrooms and false truffles. McIlvainea. 1994;11:61–74.

16. Aime MC, Miller Jr OK. Proposal to conserve the name Chroogomphusagainst Brauniellula (Gomphidiaceae, Agaricales, Basidiomycota). Taxon.2006;55:228–9.

17. Lebel T, Tonkin JE. Australasian species of Macowanites are sequestratespecies of Russula (Russulaceae, Basidiomycota). Austral Syst Bot. 2007;20:355–81.

18. Lebel T, Syme A. Sequestrate species of Agaricus and Macrolepiota fromAustralia: new combinations and species, and their position in a calibratedphylogeny. Mycologia. 2012;104:496–520.

19. Ge ZW, Smith ME. Phylogenetic analysis of rDNA sequences indicates thatthe sequestrate Amogaster viridiglebus is derived from within the agaricoidgenus Lepiota (Agaricaceae). Mycol Prog. 2013;12:151–5.

20. Pegler DN, Young TWH. The gasteroid Russulales. T Brit Mycol Soc. 1979;72:353–88.

21. Hopple Jr JS, Vilgalys R. Phylogenetic relationships in the mushroom genusCoprinus and dark-spored allies based on sequence data from the nucleargene coding for the large ribosomal subunit RNA: divergent domains,outgroups, and monophyly. Mol Phylogenet Evol. 1999;13:1–19.

22. Peintner U, Bougher NL, Castellano MA, Moncalvo J-M, Moser MM, TrappeJM, Vilgalys R. Multiple origins of sequestrate fungi related to Cortinarius(Cortinariaceae). Am J Bot. 2001;88:2168–79.

23. Justo A, Morgenstern I, Hallen-Adams HE, Hibbett DS. Convergent evolutionof sequestrate forms in Amanita under Mediterranean climate conditions.Mycologia. 2010;102:675–88.

24. Wilson AW, Binder M, Hibbett DS. Diversity and evolution ofectomycorrhizal host associations in the Sclerodermatineae (Boletales,Basidiomycota). New Phytol. 2012;194:1079–95.

25. Braaten CC, Matheny PB, Viess DL, Wood MG, Williams JH, Bougher NL. Twonew species of Inocybe from Australia and North America that include novelsecotioid forms. Botany. 2014;92:9–22.

26. Smith ME, Amese KR, Elliott TF, Obase K, Aime MC, Henkel TW. Newsequestrate fungi from Guyana: Jimtrappea guyanensis gen. sp. nov.,Castellanea pakaraimophila gen. sp. nov., and Costatisporus cyanescens gen.sp. nov. (Boletaceae, Boletales). IMA Fungus. 2015;6:297–317.

27. Henkel TW, Aime MC, Smith ME. Guyanagaster, a new wood-decayingsequestrate fungal genus related to Armillaria (Physalacriaceae, Agaricales,Basidiomycota). Am J Bot. 2010;97:1474–84.

28. Isaacs R, Gillman MP, Johnston M, Marsh F, Wood BC. Size structure of adominant Neotropical forest tree species, Dicymbe altsonii, in Guyana andsome factors reducing seedling leaf area. J Trop Ecol. 1996;12:599–606.

29. Henkel TW. Mondominance in the ectomycorrhizal Dicymbe corymbosa(Caesalpiniaceae) from Guyana. J Trop Ecol. 2003;19:417–37.

30. Zagt RJ. Pre-dispersal and early post-dispersal demography, andreproductive litter production, in the tropical tree Dicymbe altsonii inGuyana. J Trop Ecol. 1997;13:511–26.

31. Kornerup A, Wanscher JH. Methuen handbook of colour. 3rd ed. London:Eyre Methuen; 1978.

32. Gardes M, Bruns TD. ITS primers with enhanced specificity forbasidiomycetes—Application to the identification of mycorrhizae and rusts.Mol Ecol. 1993;2:113–8.

33. Vilgalys R, Hester M. Rapid genetic identification and mapping ofenzymatically amplified ribosomal DNA from several Cryptococcus species.J Bacteriol. 1990;172:4238–46.

34. Moncalvo JM, Lutzoni FM, Rehner SA, Johnson J, Vilgalys R. Phylogeneticrelationships of agaric fungi based on nuclear large subunit ribosomal DNAsequences. Syst Biol. 2000;49:278–305.

35. Rehner SA, Buckley E. A Beauveria phylogeny inferred from nuclear ITS andEF1-α sequences: evidence for cryptic diversification and links to Cordycepsteleomorphs. Mycologia. 2005;97:84–98.

36. Matheny PB. Improving phylogenetic inference of mushrooms using RPB1 andRPB2 sequences (Inocybe, Agaricales). Mol Phylogenet Evol. 2005;35:1–20.

37. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. MEGA5:molecular evolutionary genetics analysis using maximum likelihood,evolutionary distance, and maximum parsimony methods. Mol Biol Evol.2011;28:2731–9.

38. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy andhigh throughput. Nucleic Acids Res. 2004;32:1792–7.

39. Lanfear R, Calcott B, Ho SYW, Guindon S. PartitionFinder: combinedselection of partitioning schemes and substitution models for phylogeneticanalyses. Mol Biol Evol. 2012;29(6):1695–701.

40. Stamatakis A. RAxML-VI-HPC: maximum likelihood-based phylogeneticanalyses with thousands of taxa and mixed models. Bioinformatics. 2006;22:2688–90.

41. Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Höhna S, et al.MrBayes 3.2: efficient Bayesian phylogenetic inference and model choiceacross a large model space. Syst Biol. 2012;61:539–42.

42. Skrede I, Engh IB, Binder M, Carlsen T, Kauserud H, Bendiksby M.Evolutionary history of Serpulaceae (Basidiomycota): molecular phylogeny,historical biogeography and evidence for a single transition of nutritionalmode. BMC Evol Biol. 2011;11:230.

43. Ryberg M, Matheny PB. Asynchronous origins of ectomycorrhizal clades ofAgaricales. Proc Biol Sci. 2012;279:2003–11.

44. Wilson AW, Hosaka K, Mueller GM. Evolution of ectomycorrhizas as a driverof diversification and biogeographic patterns in the model mycorrhizalmushroom genus Laccaria. New Phytol. 2016;doi:10.1111/nph.14270.

45. Drummond AJ, Suchard MA, Xie D, Rambaut A. Bayesian phylogenetics withBEAUTi and the BEAST 1.7. Mol Biol Evol. 2012;29:1969–73.

46. Rambaut A, Suchard MA, Xie D, Drummond AJ. Tracer v1.6. 2014. http://beast.bio.ed.ac.uk/Tracer. Accessed 6 June 2016.

47. Jenkinson TS, Perry BA, Schaefer RE, Desjardin DE. Cryptomarasmius gen.nov. established in the Physalacriaceae to accommodate members ofMarasmius section Hygrometrici. Mycologia. 2014;106(1):86–94.

48. Dentinger BTM, Gaya E, O’Brien H, Suz LM, Lachlan R, Díaz-Valderrama JR, et al.Tales from the crypt: genome mining from fungarium specimens improvesresolution of the mushroom tree of life. Biol J Linn Soc. 2016;117:11–32.

49. Hibbett DS, Grimaldi D, Donoghue MJ. Fossil mushrooms from Mioceneand Cretaceous ambers and the evolution of Homobasidiomycetes. Am JBot. 1997;84(8):981–91.

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 14 of 16

http://dx.doi.org/10.1111/nph.14270http://beast.bio.ed.ac.uk/Tracerhttp://beast.bio.ed.ac.uk/Tracer

50. Ree RH, Smith SE. Maximum likelihood inference of geographic range evolutionby dispersal, local extinction, and cladogenesis. Syst Biol. 2008;57:4–14.

51. Yu Y, Harris AJ, Blair C, He X. RASP (Reconstruct Ancestral State inPhylogenies): a tool for historical biogeography. Mol Phylogenet Evol. 2015;87:46–9.

52. Singer R. Flora Neotropica monograph 3, Omphalinae. New York: HafnerPublishing; 1970.

53. Hennings P. Fungi camerunenses. I Bot Jb. 1895;22:70–111.54. Watling R. Armillaria Staude in the Cameroon Republic. Persoonia. 1992;14:

483–91.55. Zachos J, Pagani M, Sloan L, Thomas E, Billups K. Trends, rhythms, and

aberrations in global climate 65 Ma to present. Science. 2001;292:686–93.56. Wolfe JA. Distributions of major vegetation types during the Tertiary. In:

Sundquist ET, Broekner WS, editors. The Carbon Cycle and Atmospheric CO2:Natural Variations, Archean to Present, American Geophysical Union Monograph32. Washington, DC: American Geophysical Union; 1985. p. 357–76.

57. Zolciak A, Bouteville R-J, Tourvieille J, Roeckel-Drevet P, Nicolas P, Guillaumin J-J.Occurrence of Armillaria ectypa (Fr.) Lamoure in peat bogs of the Auvergne—thereproduction system of the species. Cryptogamie Mycol. 1997;18(4):299–313.

58. Guillaumin J-J, Mohammed C, Anselmi N, Courtecuisse R, Gregory SC,Holdenrieder O, et al. Geographical distribution and ecology of theArmillaria species in western Europe. Eur J Forest Pathol. 1993;23:321–41.

59. Volk TJ, Burdsall Jr HH. A nomenclatural study of Armillaria and Armillariellaspecies. Synopsis Fungorum 8. Oslo: Fungiflora; 1995.

60. Antonín V, Jankovský L, Lochman J, Tomšovský M. Armillariasocialis—morphological-anatomical and ecological characteristics,pathology, distribution in the Czech Republic and Europe and remarks onits genetic variation. Czech Mycol. 2006;58(3–4):209–24.

61. Cassens I, Vicario S, Waddell VG, Balchowsky H, Van Belle D, Ding W, et al.Independent adaptation to riverine habitats allowed survival of ancientcetacean lineages. Proc Natl Acad Sci U S A. 2000;97(21):11343–7.

62. Morley RJ. Cretaceous and Tertiary climate change and the past distributionof megathermal rainforests. In: Bush MB, Flenley JR, Gosling WD, editors.Tropical rainforest responses to climatic change. Heidelberg: Springer Praxis;2011. p. 1–34.

63. Tiffney BH. Perspectives on the origin of the floristic similarity betweeneastern Asia and Eastern North America. J Arnold Arboretum. 1985;66:73–94.

64. Maphosa L, Wingfield BD, Coetzee MPA, Mwenje E, Wingfield MJ. Phylogeneticrelationships among Armillaria species inferred from partial elongation factor1-alpha DNA sequence data. Australas Plant Pathol. 2006;35:513–20.

65. Dercourt J, Zonenshain LP, Ricou L-E, Kazmin VG, Le Pichon X, Knipper AL,et al. Geological evolution of the Tethys belt from the Atlantic to the Pamirssince the Lias. Tectonophysics. 1986;123:241–315.

66. Coetzee MPA, Wingfield BD, Harrington TC, Dalevi D, Coutinho TA,Wingfield MJ. Geographic diversity of Armillaria mellea s.s. based onphylogenetic analysis. Mycologia. 2000;92:105–13.

67. Crayn DM, Costion C, Harrington MG. The Sahul-Sunda floristic exchange:dated molecular phylogenies document Cenozoic intercontinental dispersaldynamics. J Biogeogr. 2015;42:11–24.

68. Coetzee MPA, Wingfield BD, Bloomer P, Ridley GS, Wingfield MJ. Molecularidentification and phylogeny of Armillaria isolates from South America andIndo-Malaysia. Mycologia. 2003;95(2):285–93.

69. Webb PN, Harwood DM. Pliocene fossil Nothofagus (Southern Beech) fromAntarctica: phytogeography, dispersal strategies, and survival in highlatitude glacial-deglacial environments. In: Alden JN, Mastrantonio JL, ØdumS, editors. Forest Development in Cold Climates, NATO ASI Series 244. NewYork: Springer US; 1993. p. 135–66.

70. Heads M. Panbiogeography of Nothofagus (Nothofagaceae): analysis of themain species massings. J Biogeogr. 2006;33:1066–75.

71. Garraway MO, Hütterman A, Wargo PM. Ontogeny and physiology. In: ShawCG, Kile GA, editors. Armillaria Root Disease. Agriculture Handbook 691.Washington DC: USDA Forest Service; 1991. p. 21–47.

72. Smith ML, Bruhn JN, Anderson JB. The fungus Armillaria bulbosa is amongthe largest and oldest living organisms. Nature. 1992;356(2):428–31.

73. Rhoads AS. A comparative study of two closely related root-rot fungi,Clitocybe tabescens and Armillaria mellea. Mycologia. 1945;61:741–66.

74. Swift MJ. The ecology of Armillaria mellea Vahl (ex Fries) in the indigenousand exotic woodlands of Rhodesia. Forestry Oxford. 1972;45:67–86.

75. Coetzee MPA, Wingfield BD, Coutinho TA, Wingfield MJ. Identification of thecausal agent of Armillaria root rot of Pinus species in South Africa.Mycologia. 2000;92:777–85.

76. Rizzo DM, Blanchette RA, Palmer MA. Biosorption of metal ions by Armillariarhizomorphs. Can J Bot. 1992;70(8):1515–20.

77. Mironenko NV, Alekhina IA, Zhdanova NN, Bulat SA. Intraspecific variation ingamma-radiation resistance and genomic structure in the filamentousfungus Alternaria alternata: a case study of strains inhabiting Chernobylreactor, 4. Ecotoxicol Environ Saf. 2000;45:177–87.

78. Mihail JD, Bruhn JN, Leininger TD. The effects of moisture and oxygenavailability on rhizomorph generation by Armillaria tabescens in comparisonwith A. gallica and A. mellea. Mycol Res. 2002;106(6):697–704.

79. Watkinson SC, Boddy L, Money NP. The Fungi. 3rd ed. Cambridge:Academic; 2015.

80. Singer R. Agaricales in modern taxonomy. Lilloa. 1951;22:5–832.81. Singer R. The Agaricales in modern taxonomy. 4th ed. Königstein: Koeltz

Scientific Books; 1986.82. Herink J. Taxonomie václacky obecné—Armillaria mellea (Vahl ex Fr.) Kumm. In:

Hasek J, editor. Sympozium o václavce obecné: Armillaria mellea (Vahl ex Fr.)Kumm. Brno, Czechoslovakia: Vysoká Skola Zemĕdĕlská v Brné; 1973. p. 21–48.

83. Brazee NJ, Ortiz-Santana B, Banik MT, Lindner DL. Armillaria altimontana, anew species from the western interior of North America. Mycologia. 2012;104(5):1200–5.

84. Gregory SC, Watling R. Occurerence of Armillaria borealis in Britain. T BritMycol Soc. 1985;84(1):47–55.

85. Bauce E, Allen DC. Role of Armillaria calvescens and Glycobius speciosus in asugar maple decline. Can J For Res. 1992;22:549–52.

86. Dessureault M. Morphological studies of the Armillaria mellea complex: Twonew species, A. gemina and A. calvescens. Mycologia. 1989;81(2):216–25.

87. Prospero S, Holdenrieder O, Rigling D. Comparison of the virulence ofArmillaria cepistipes and Armillaria ostoyae on four Norway spruceprovenances. Forest Pathol. 2004;24(1):1–14.

88. Mwenje E, Ride JP, Pearce RB. Distribution of Zimbabwean Armillaria groupsand their pathogenicity on cassava. Plant Pathol. 1998;47:623–34.

89. Kile GA. Genotypes of Armillaria hinnulea in wet sclerophyll eucalypt forestin Tasmania. Trans Br Mycol Soc. 1986;87(2):312–4.

90. Ramsfield TD, Power MWP, Ridley GS. A comparison of populations ofArmillaria hinnulea in New Zealand and Australia. N Z Plant Prot. 2008;61:41–7.

91. Shaw III CG, MacKenzie M, Toes EHA, Hood IA. Cultural characteristics andpathogenicity to Pinus radiata of Armillaria novae-zelandiae and A. limonea.N Z J For Sci. 1981;11(1):65–70.

92. Shaw III CG, Calderon S. Impact of Armillaria root rot in plantations of Pinusradiata established on sites converted from indigenous forest. N Z J For Sci.1977;7:359–73.

93. Kile GA. Armillaria luteobubalina—a primary cause of decline and death oftrees in mixed species eucalypt forests in central Victoria. Aust Forest Res.1981;11(1):63–7.

94. Pareek M, Cole L, Ashford AE. Variations in structure of aerial andsubmerged rhizomorphs of Armillaria luteobubalina indicate that they maybe organs of adsorption. Mycol Res. 2001;105:1377–87.

95. Pildain MB, Coetzee MPA, Wingfield BD, Wingfield MJ, Rajchenberg M.Taxonomy of Armillaria in the Patagonian forests of Argentina. Mycologia.2010;102(2):392–403.

96. Raabe RD. Host list of the root rot fungus Armillaria mellea. Hilgardia. 1962;32:25–88.

97. Cleary MR, van der Kamp BJ, Morrison DJ. Pathogenicity and virulence ofArmillaria sinapina and host response to infection in Douglas-fir, westernhemlock and western red cedar in the southern Interior of British Columbia.Forest Pathol. 2012;42(6):481–91.

98. Burdsall Jr HH, Volk TJ. Armillaria solidipes, an older name for the funguscalled Armillaria ostoyae. N Am Fungi. 2008;3(7):261–7.

99. Moncalvo J-M, Vilgalys R, Redhead SA, Johnson JE, James TY, Aime MC,et al. One hundred and seventeen clades of eugarics. Mol Phylogenet Evol.2002;23:357–400.

100. Brazee NJ, Hulvey JP, Wick RL. Evaluation of partial tef1, rpb2, and nLSUsequences for identification of isolates representing Armillaria calvescensand Armillaria gallica from northeastern North America. Fungal Biol. 2011;115(8):741–9.

101. Guo T, Wang HC, Xue WQ, Zhao J, Yang ZL. Phylogenetic analyses ofArmillaria reveal at least 15 phylogenetic lineages in China, seven of whichare associated with cultivated Gastrodia elata. PLoS One. 2016;11(5):e0154794.

102. Wingfield BD, Ambler JM, Coetzee MPA, de Beer ZW, Duong TA, Joubert F,et al. Draft genome sequences of Armillaria fuscipes, Certocystiopsis minuta,

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 15 of 16

Ceratocystis adiposa, Endoconidiophora laricicola, E. polonica and Penicilliumfreii DAOMC 242723. IMA Fungus. 2016;7(1):217–27.

103. Binder M, Hibbett DS, Wang Z, Farnham WF. Evolutionary relationships ofMycaureola dilseae (Agaricales), a basidiomycete pathogen of subtidalrhodophyte. Am J Bot. 2006;93:547–56.

104. Matheny PB, Wang Z, Binder M, Curtis JM, Lim YW, Nilsson RH, et al.Contributions of rpb2 and tef1 to the phylogeny of mushrooms and allies(Basidiomycota, Fungi). Mol Phylogenet Evol. 2007;43:430–51.

105. Assembling the Fungal Tree of Life. 1995. www.aftol.org. Accessed 6 June 2016.106. Petersen RH, Hughes KW. The Xerula/Oudemansiella Complex (Agaricales).

Nova Hedwigia Beih. 2010;137:625.107. Grigoriev IV, Nikitin R, Haridas S, Kuo A, Ohm R, Otillar R, et al. MycoCosm portal:

gearing up for 1000 fungal genomes. Nucl Acids Res. 2013;42:D699–D704.108. Hibbett DS, Gilbert L-B, Donoghue MJ. Evolutionary instability of

ectomycorrhizal symbioses in basidiomycetes. Nature. 2000;407:506–8.109. Wang Z, Binder M, Dai YC, Hibbett DS. Phylogenetic relationships of

Sparassis inferred from nuclear and mitochondrial ribosomal DNA and RNApolymerase sequences. Mycologia. 2004;96(5):1015–29.

110. Matheny PB, Curtis JM, Hofstetter V, Aime MC, Moncalvo JM, Ge ZW, et al.Major clades of Agaricales: a multilocus phylogenetic overview. Mycologia.2006;98(6):982–95.

111. Binder M, Larsson KH, Matheny PB, Hibbett DS. Amylocorticiales ord. nov.and Jaapiales ord. nov.: early diverging clades of Agaricomycetidaedominated by corticioid forms. Mycologia. 2010;102(4):865–80.

112. Ronikier M, Ronikier A. Rhizomarasmius epidryas (Physalacriaceae):phylogenetic placement of an arctic-alpine fungus with obligate saprobicaffinity to Dryas spp. Mycologia. 2011;103(5):1124–32.

113. Qin J, Hao Y-J, Yang ZL, Li Y-C. Paraxerula ellipsospora, a new Asian speciesof Physalacriaceae. Mycol Prog. 2014;13(3):639–47.

114. Qin J, Yang ZL. Three new species of Physalacria from China, with a key tothe Asian taxa. Mycologia. 2016;108(1):215–26.

• We accept pre-submission inquiries • Our selector tool helps you to find the most relevant journal• We provide round the clock customer support • Convenient online submission• Thorough peer review• Inclusion in PubMed and all major indexing services • Maximum visibility for your research

Submit your manuscript atwww.biomedcentral.com/submit

Submit your next manuscript to BioMed Central and we will help you at every step:

Koch et al. BMC Evolutionary Biology (2017) 17:33 Page 16 of 16

http://www.aftol.org/

AbstractBackgroundResultsConclusions

BackgroundMethodsCollecting and morphological analysesMolecular methodsPhylogenetic analysesDivergence time estimationAncestral range estimation

ResultsSequences, collections and culturesResolved Guyanagaster phylogenetic hypothesisAge and ancestral range of Armillaria and Guyanagaster

DiscussionBiogeography and evolution of Guyanagaster and DesarmillariaBiogeography and spread of ArmillariaThe rise of rhizomorphs: morphological developments in the armillarioid clade

ConclusionsFormal taxonomic descriptions

Additional filesAcknowledgementsFundingAvailability of data and materialsAuthors’ contributionsCompeting interestsAuthor detailsReferences